Lung CT Densitometry in Idiopathic Pulmonary Fibrosis for the Prediction of Natural Course, Severity, and Mortality

Deep learning-based prognostication in idiopathic pulmonary fibrosis using chest radiographs

OBJECTIVES

To develop and validate a deep learning-based prognostic model in patients with idiopathic pulmonary fibrosis (IPF) using chest radiographs.

METHODS

To develop a deep learning-based prognostic model using chest radiographs (DLPM), the patients diagnosed with IPF during 2011-2021 were retrospectively collected and were divided into training (n = 1007), validation (n = 117), and internal test (n = 187) datasets. Up to 10 consecutive radiographs were included for each patient. For external testing, three cohorts from independent institutions were collected (n = 152, 141, and 207). The discrimination performance of DLPM was evaluated using areas under the time-dependent receiver operating characteristic curves (TD-AUCs) for 3-year survival and compared with that of forced vital capacity (FVC). Multivariable Cox regression was performed to investigate whether the DLPM was an independent prognostic factor from FVC. We devised a modified gender-age-physiology (GAP) index (GAP-CR), by replacing DLCO with DLPM.

RESULTS

DLPM showed similar-to-higher performance at predicting 3-year survival than FVC in three external test cohorts (TD-AUC: 0.83 [95% CI: 0.76-0.90] vs. 0.68 [0.59-0.77], p < 0.001; 0.76 [0.68-0.85] vs. 0.70 [0.60-0.80], p = 0.21; 0.79 [0.72-0.86] vs. 0.76 [0.69-0.83], p = 0.41). DLPM worked as an independent prognostic factor from FVC in all three cohorts (ps < 0.001). The GAP-CR index showed a higher 3-year TD-AUC than the original GAP index in two of the three external test cohorts (TD-AUC: 0.85 [0.80-0.91] vs. 0.79 [0.72-0.86], p = 0.02; 0.72 [0.64-0.80] vs. 0.69 [0.61-0.78], p = 0.56; 0.76 [0.69-0.83] vs. 0.68 [0.60-0.76], p = 0.01).

CONCLUSIONS

A deep learning model successfully predicted survival in patients with IPF from chest radiographs, comparable to and independent of FVC.

CLINICAL RELEVANCE STATEMENT

Deep learning-based prognostication from chest radiographs offers comparable-to-higher prognostic performance than forced vital capacity.

KEY POINTS

• A deep learning-based prognostic model for idiopathic pulmonary fibrosis was developed using 6063 radiographs. • The prognostic performance of the model was comparable-to-higher than forced vital capacity, and was independent from FVC in all three external test cohorts. • A modified gender-age-physiology index replacing diffusing capacity for carbon monoxide with the deep learning model showed higher performance than the original index in two external test cohorts.

Establishment and application of the BRP prognosis model for idiopathic pulmonary fibrosis

BACKGROUND

Idiopathic pulmonary fibrosis (IPF) is the most common idiopathic interstitial lung disease. Clinical models to accurately evaluate the prognosis of IPF are currently lacking. This study aimed to construct an easy-to-use and robust prediction model for transplant-free survival (TFS) of IPF based on clinical and radiological information.

METHODS

A multicenter prognostic study was conducted involving 166 IPF patients who were followed up for 3 years. The end point of follow-up was death or lung transplantation. Clinical information, lung function tests, and chest computed tomography (CT) scans were collected. Body composition quantification on CT was performed using 3D Slicer software. Risk factors in blood routine examination-radiology-pulmonary function (BRP) were identified by Cox regression and utilized to construct the "BRP Prognosis Model". The performance of the BRP model and the gender-age-physiology variables (GAP) model was compared using time-ROC curves, calibration curves, and decision curve analysis (DCA). Furthermore, histopathology fibrosis scores in clinical specimens were compared between the different risk stratifications identified by the BRP model. The correlations among body composition, lung function, serum inflammatory factors, and profibrotic factors were analyzed.

RESULTS

Neutrophil percentage > 68.3%, pericardial adipose tissue (PAT) > 94.91 cm3, pectoralis muscle radiodensity (PMD) ≤ 36.24 HU, diffusing capacity of the lung for carbon monoxide/alveolar ventilation (DLCO/VA) ≤ 56.03%, and maximum vital capacity (VCmax) < 90.5% were identified as independent risk factors for poor TFS among patients with IPF. We constructed a BRP model, which showed superior accuracy, discrimination, and clinical practicability to the GAP model. Median TFS differed significantly among patients at different risk levels identified by the BRP model (low risk: TFS > 3 years; intermediate risk: TFS = 2-3 years; high risk: TFS ≈ 1 year). Patients with a high-risk stratification according to the BRP model had a higher fibrosis score on histopathology. Additionally, serum proinflammatory markers were positively correlated with visceral fat volume and infiltration.

CONCLUSIONS

In this study, the BRP prognostic model of IPF was successfully constructed and validated. Compared with the commonly used GAP model, the BRP model had better performance and generalization with easily obtainable indicators. The BRP model is suitable for clinical promotion.

Histogram-based analysis in progressive pulmonary fibrosis: relationships between pulmonary functional tests and HRCT indexes

OBJECTIVES

To investigate relationships between histogram-based high-resolution CT (HRCT) indexes and pulmonary function tests (PFTs) in interstitial lung diseases.

METHODS

Forty-nine patients having baseline and 1-year HRCT examinations and PFTs were investigated. Histogram-based HRCT indexes were calculated; strength of associations with PFTs was investigated using Pearson correlation. Patients were divided into progressive and non-progressive groups. HRCT indexes were compared between the two groups using the U-test; within each group, baseline and follow-up Wilcoxon analysis was performed. Receiver operating characteristic analysis was used for predicting disease progression.

RESULTS

At baseline, moderate correlations were observed considering kurtosis and diffusion capacity of the lungs for carbon monoxide (DLCO) (r = 0.54) and skewness and DLCO (r = 0.559), whereas weak but significant correlations were observed between forced vital capacity and kurtosis (r = 0.368, p = 0.009) and forced vital capacity and skewness (r = 0.391, p = 0.005). Negative correlations were reported between HAA% and PFTs (from r = -0.418 up to r = -0.507). At follow-up correlations between quantitative indexes and PFTs were also moderate, except for high attenuation area (HAA)% -700 and DLCO (r = -0.397). In progressive subgroup, moderate and strong correlations were found between DLCO and HRCT indexes (r = 0.595 kurtosis, r = 0.672 skewness, r=-0. 598 HAA% -600 and r = -0.626 HAA% -700). At follow-up, we observed significant differences between the two groups for kurtosis (p = 0.029), HAA% -600 (p = 0.04) and HAA% -700 (p = 0.02). To predict progression, ROC analysis reported sensitivity of 90.9% and specificity of 51.9% using a threshold value of δ kurtosis <0.03.

CONCLUSION

At one year, moderate correlations suggest that progression could be assessed through HRCT quantification.

ADVANCES IN KNOWLEDGE

This study promotes histogram-based HRCT indexes in the assessment of progressive pulmonary fibrosis.

Diagnostic finding on high-resolution computed tomography (HRCT) predicts a good response to pirfenidone in patients with idiopathic pulmonary fibrosis

Idiopathic pulmonary fibrosis (IPF) is a debilitating condition, with a life expectancy of 2 to 5 years after diagnosis. Pirfenidone is a drug that has been shown to reduce the decline in forced vital capacity (FVC). We sought to identify whether different patterns on high-resolution computed tomography (HRCT) have different clinical effects through a retrospective comparison of baseline values and changes in pulmonary function tests (PFTs) after treatment with pirfenidone. We retrospectively analyzed data from IPF patients treated with pirfenidone at Nanjing Drum Tower Hospital in Jiangsu Province, China. According to the HRCT pattern, the patients were divided into usual interstitial pneumonitis (UIP) and possible UIP groups. Baseline clinical characteristics and changes every 6 months in the PFTs during the follow-up period were compared between the 2 groups. A total of 65 consecutive patients were enrolled. According to the HRCT pattern, patients were clustered into the UIP group (n = 46) and possible UIP group (n = 19). No difference was observed in the baseline PFTs ratio between the 2 groups. The FVC values of the 2 groups were not significantly different at the initial treatment and at 6 and 12 months after pirfenidone treatment (P = .081, 0.099, and 0.236, respectively). The improvement in % diffusion capacity of the lung for carbon monoxide (%DLCO) was higher in the possible UIP group after 6 and 12 months of pirfenidone treatment (P = .149, 0.026, and 0.025, respectively). The annual decrease in FVC was not significantly different between the 2 groups, and the annual decrease in %DLCO in the UIP group was significantly higher than that in patients with the possible UIP type (-7.767 ± 12.797 vs 0.342 ± 20.358, P < .05). These results indicate that patients with IPF with a possible UIP pattern on HRCT showed indications of a good response to pirfenidone.

Prognostic value of deep learning-based fibrosis quantification on chest CT in idiopathic pulmonary fibrosis

OBJECTIVE

To investigate the prognostic value of deep learning (DL)-driven CT fibrosis quantification in idiopathic pulmonary fibrosis (IPF).

METHODS

Patients diagnosed with IPF who underwent nonenhanced chest CT and spirometry between 2005 and 2009 were retrospectively collected. Proportions of normal (CT-Norm%) and fibrotic lung volume (CT-Fib%) were calculated on CT using the DL software. The correlations of CT-Norm% and CT-Fib% with forced vital capacity (FVC) and diffusion capacity of carbon monoxide (D_LCO) were evaluated. The multivariable-adjusted hazard ratios (HRs) of CT-Norm% and CT-Fib% for overall survival were calculated with clinical and physiologic variables as covariates using Cox regression. The feasibility of substituting CT-Norm% for D_LCO in the GAP index was investigated using time-dependent areas under the receiver operating characteristic curve (TD-AUCs) at 3 years.

RESULTS

In total, 161 patients (median age [IQR], 68 [62-73] years; 104 men) were evaluated. CT-Norm% and CT-Fib% showed significant correlations with FVC (Pearson's r, 0.40 for CT-Norm% and - 0.37 for CT-Fib%; both p < 0.001) and D_LCO (0.52 for CT-Norm% and - 0.46 for CT-Fib%; both p < 0.001). On multivariable Cox regression, both CT-Norm% and CT-Fib% were independent prognostic factors when adjusted to age, sex, smoking status, comorbid chronic diseases, FVC, and D_LCO (HRs, 0.98 [95% CI 0.97-0.99; p < 0.001] for CT-Norm% at 3 years and 1.03 [1.01-1.05; p = 0.01] for CT-Fib%). Substituting CT-Norm% for D_LCO showed comparable discrimination to the original GAP index (TD-AUC, 0.82 [0.78-0.85] vs. 0.82 [0.79-0.86]; p = 0.75).

CONCLUSION

CT-Norm% and CT-Fib% calculated using chest CT-based deep learning software were independent prognostic factors for overall survival in IPF.

KEY POINTS

• Normal and fibrotic lung volume proportions were automatically calculated using commercial deep learning software from chest CT taken from 161 patients diagnosed with idiopathic pulmonary fibrosis. • CT-quantified volumetric parameters from commercial deep learning software were correlated with forced vital capacity (Pearson's r, 0.40 for normal and - 0.37 for fibrotic lung volume proportions) and diffusion capacity of carbon monoxide (Pearson's r, 0.52 and - 0.46, respectively). • Normal and fibrotic lung volume proportions (hazard ratios, 0.98 and 1.04; both p < 0.001) independently predicted overall survival when adjusted for clinical and physiologic variables.

Quantitative computed tomography predicts outcomes in idiopathic pulmonary fibrosis

BACKGROUND AND OBJECTIVE

Prediction of disease course in patients with progressive pulmonary fibrosis remains challenging. The purpose of this study was to assess the prognostic value of lung fibrosis extent quantified at computed tomography (CT) using data-driven texture analysis (DTA) in a large cohort of well-characterized patients with idiopathic pulmonary fibrosis (IPF) enrolled in a national registry.

METHODS

This retrospective analysis included participants in the Australian IPF Registry with available CT between 2007 and 2016. CT scans were analysed using the DTA method to quantify the extent of lung fibrosis. Demographics, longitudinal pulmonary function and quantitative CT metrics were compared using descriptive statistics. Linear mixed models, and Cox analyses adjusted for age, gender, BMI, smoking history and treatment with anti-fibrotics were performed to assess the relationships between baseline DTA, pulmonary function metrics and outcomes.

RESULTS

CT scans of 393 participants were analysed, 221 of which had available pulmonary function testing obtained within 90 days of CT. Linear mixed-effect modelling showed that baseline DTA score was significantly associated with annual rate of decline in forced vital capacity and diffusing capacity of carbon monoxide. In multivariable Cox proportional hazard models, greater extent of lung fibrosis was associated with poorer transplant-free survival (hazard ratio [HR] 1.20, p < 0.0001) and progression-free survival (HR 1.14, p < 0.0001).

CONCLUSION

In a multi-centre observational registry of patients with IPF, the extent of fibrotic abnormality on baseline CT quantified using DTA is associated with outcomes independent of pulmonary function.

Heterogeneity of Lung Density in Patients With Chronic Thromboembolic Pulmonary Hypertension (CTEPH)

RATIONALE AND OBJECTIVES

Pulmonary endarterectomy (PEA) is one of the most effective treatments for chronic thromboembolic pulmonary hypertension (CTEPH). Right heart catheterization (RHC) is the gold standard assessment for pulmonary circulatory dynamics. However, computed tomography (CT) is less invasive than RHC and can elucidate some of the morphological changes caused by thromboembolism. We hypothesized that CT could facilitate the evaluation of heterogeneous pulmonary perfusion. This study investigated whether CT imaging features reflect the disease severity and changes in pulmonary circulatory dynamics in patients with CTEPH before and after PEA.

MATERIALS AND METHODS

This retrospective study included 58 patients with CTEPH who underwent PEA. Pre-PEA and post-PEA CT images were assessed for heterogeneity using CT texture analysis (CTTA). The CT parameters were compared with the results of the RHC and other clinical indices and analyzed with receiver operating characteristic curves analysis for patients with and without residual pulmonary hypertension (PH) (post-PEA mean pulmonary artery pressure ≥ 25 mmHg).

RESULTS

CT measurements reflecting heterogeneity were significantly correlated with mean pulmonary artery pressure. Kurtosis, skewness, and uniformity were significantly lower, and entropy was significantly higher in patients with residual PH than patients without residual PH. Area under the curve values of pre-PEA and post-PEA entropy between patients with and without residual PH were 0.71 (95% confidence interval 0.57-0.84) and 0.75 (0.63-0.88), respectively.

CONCLUSION

Heterogeneity of lung density might reflect pulmonary circulatory dynamics, and CTTA for heterogeneity could be a less invasive technique for evaluation of changes in pulmonary circulatory dynamics in patients with CTEPH undergoing PEA.

Quantitative Computed Tomography: What Clinical Questions Can it Answer in Chronic Lung Disease?

Quantitative computed tomography (QCT) has recently gained an important role in the functional assessment of chronic lung disease. Its capacity in diagnostic, staging, and prognostic evaluation in this setting is similar to that of traditional pulmonary function testing. Furthermore, it can demonstrate lung injury before the alteration of pulmonary function test parameters, and it enables the classification of disease phenotypes, contributing to the customization of therapy and performance of comparative studies without the intra- and inter-observer variation that occurs with qualitative analysis. In this review, we address technical issues with QCT analysis and demonstrate the ability of this modality to answer clinical questions encountered in daily practice in the management of patients with chronic lung disease.

The Normal Lung Index From Quantitative Computed Tomography for the Evaluation of Obstructive and Restrictive Lung Disease

PURPOSE

Our objective was to evaluate whether the normal lung index (NLI) from quantitative computed tomography (QCT) analysis can be used to predict mortality as well as pulmonary function tests (PFTs) in patients with chronic obstructive pulmonary disease (COPD) and interstitial lung disease (ILD).

MATERIALS AND METHODS

Normal subjects (n=20) and patients with COPD (n=172) and ILD (n=114) who underwent PFTs and chest CT were enrolled retrospectively in this study. QCT measures included the NLI, defined as the ratio of the lung with attenuation between -950 and -700 Hounsfield units (HU) over the total lung volume (-1024 to -250 HU, mL), high-attenuation area (-700 to -250 HU, %), emphysema index (>6% of pixels < -950 HU), skewness, kurtosis, and mean lung attenuation. Coefficients of correlation between QCT measurements and PFT results in all subjects were calculated. Univariate and multivariate survival analyses were performed to assess mortality prediction by disease.

RESULTS

The Pearson correlation analysis showed that the NLI correlated moderately with the forced expiratory volume in 1 second in subjects with COPD (r=0.490, P<0.001) and the forced vital capacity in subjects with ILD (r=0.452, P<0.001). Multivariate analysis revealed that the NLI of <70% was a significant independent predictor of mortality in subjects with COPD (hazard ratio=3.14, P=0.034) and ILD (hazard ratio=2.72, P=0.005).

CONCLUSION

QCT analysis, specifically the NLI, can also be used to predict mortality in individuals with COPD and ILD.

Novel Artificial Intelligence-based Technology for Chest Computed Tomography Analysis of Idiopathic Pulmonary Fibrosis

RATIONALE

There is a growing need to accurately estimate the prognosis of idiopathic pulmonary fibrosis (IPF) in clinical practice, given the development of effective drugs for treating IPF.

OBJECTIVE

To develop artificial intelligence-based image analysis software to detect parenchymal and airway abnormalities on chest computed tomography (CT) and to explore their prognostic importance in patients with IPF.

METHODS

A novel artificial intelligence-based quantitative CT image analysis software (AIQCT) was developed by applying 304 HRCT scans from patients with diffuse lung diseases as the training set. AIQCT automatically categorized and quantified ten types of parenchymal patterns as well as airways, expressing the volumes as percentages of the total lung volume. To validate the software, the area percentages of each lesion quantified by AIQCT were compared with those of the visual scores using 30 plain HRCT images with lung diseases. In addition, three-dimensional analysis for similarity with ground truth was performed using HRCT images from 10　patients with IPF. AIQCT was then applied to 120 patients with IPF who underwent chest HRCT scanning at our institute. Associations between the measured volumes and survival were analyzed.

RESULTS

The correlations between AIQCT and the visual scores were moderate to strong (correlation coefficient 0.44 to 0.95) depending on the parenchymal pattern. The Dice indexes for similarity between AIQCT data and ground truth were 0.67, 0.76, and 0.64 for reticulation, honeycomb, and bronchi, respectively. During a median follow-up period of 2,184 days, 66 patients died, and 1 underwent lung transplantation. In multivariable Cox regression analysis, bronchial volumes [adjusted hazard ratio (HR), 1.33; 95% confidence interval (CI), 1.16 to 1.53] and normal lung volumes (adjusted HR, 0.97; 95% CI, 0.94 to 0.99) were independently associated with survival after adjusting for the GAP stage of IPF.

CONCLUSIONS

Our newly developed artificial intelligence-based image analysis software successfully quantified parenchymal lesions and airway volumes. Bronchial and normal lung volumes on chest HRCT may provide additional prognostic information on the GAP stage of IPF.

NEDD4L-induced β-catenin ubiquitination suppresses the formation and progression of interstitial pulmonary fibrosis via inhibiting the CTHRC1/HIF-1α axis

Interstitial pulmonary fibrosis (IPF) is a severe progressive lung disease with limited therapeutic options and poor prognosis. Initially, we found the downregulated level of neural precursor cell expressed developmentally down-regulated 4-like protein (NEDD4L) in IPF-related expression microarray dataset, and this study was thus performed to explore the molecular mechanism of NEDD4L in IPF. The expression of NEDD4L was subsequently validated in lung tissues of IPF patients and mouse models. Then, mouse primary lung fibroblasts (LFs) were collected for in vitro functional experiments, with CCK-8, Transwell, and immunofluorescence assays used to examine the viability, migration, and differentiation of LFs. The in vitro findings were further assessed using in vivo mouse models. The expression of NEDD4L was down-regulated in lung tissues of IPF patients and mouse models. Overexpression of NEDD4L restricted the formation and progression of IPF in mice and attenuated the proliferative, invasive and differentiative abilities of LFs. Further, NEDD4L halted LFs activity by enhancing β-catenin ubiquitination and down-regulating the CTHRC1/HIF-1α axis. Also, in vivo experiments then validated that NEDD4L silencing repressed β-catenin ubiquitination and activated the CTHRC1/HIF-1α axis, thereby aggravating IPF in mice. NEDD4L may suppress the formation and progression of IPF through augmenting β-catenin ubiquitination and inhibiting the CTHRC1/HIF-1α axis.

Cutoff Points for Step Count to Predict 1-year All-Cause Mortality in Patients with Idiopathic Pulmonary Fibrosis

BACKGROUND

Although physical activity is associated with mortality in patients with idiopathic pulmonary fibrosis (IPF), reference values to interpret levels of physical activity are lacking.

OBJECTIVES

This study aimed to investigate the prognostic significance of physical activity assessed by step count and its cutoff points for all-cause mortality.

METHODS

We measured physical activity (steps per day) using an accelerometer in patients with IPF at the time of diagnosis. Relationships among physical activity and mortality, as well as cutoff points of daily step count to predict all-cause mortality were examined.

RESULTS

Eighty-seven patients (73 males) were enrolled. Forty-four patients (50.1%) died during the follow-up (median 54 months). In analysis adjusting for Gender-Age-Physiology stage and 6-min walk distance, daily step count was an independent predictor of all-cause mortality (hazard ratio (HR) = 0.820, 95% confidence interval (CI) = 0.694-0.968, p = 0.019). The optimal cutoff point (receiving operating characteristic analysis) for 1-year mortality was 3,473 steps per day (sensitivity = 0.818 and specificity = 0.724). Mortality was significantly lower in patients with a daily step count exceeding 3,473 steps than in those whose count was 3,473 or less (HR = 0.395, 95% CI = 0.218-0.715, p = 0.002).

CONCLUSIONS

Step count, an easily interpretable measurement, was a significant predictor of all-cause mortality in patients with IPF. At the time of diagnosis, a count that exceeded the cutoff point of 3,473 steps/day more than halved mortality. These findings highlight the importance of assessing physical activity in this patient population.

Utility of fractional exhaled nitric oxide in interstitial lung disease

The majority of interstitial lung diseases (ILDs) develop rapidly and are associated with a poor prognosis. Therefore, new noninvasive markers are needed to guide the classification and prognostication of ILD. We enrolled 95 patients with ILD, including dermatomyositis-associated ILD (n =69), Sjögren's syndrome-associated ILD (n= 7), mixed connective tissue disease-associated ILD (n= 9), idiopathic pulmonary fibrosis (n= 5) and hypersensitivity pneumonitis (n= 5), 82 patients with connective tissue disease but without ILD as well as 24 healthy controls, then evaluated fractional exhaled nitric oxide (FeNO50; 50 ml s^-1) (Bisenkovet al2006Vestn. Khir. Im. I. I. Grek.1659-14), pulmonary function and high-resolution computed tomography (HRCT) scores. Blood samples were analyzed and bronchoalveolar lavage fluid parameters were measured. There was no significant difference in FeNO50 values between different subgroups of ILD patients or between different subgroups of ILD patients and healthy controls. However, we found that FeNO50 was negatively correlated with the HRCT score and positively correlated with forced vital capacity. FeNO50 values did not play a clinical role in the diagnosis, differential diagnosis or prognostication of ILD.

Robust prediction of mortality of COVID-19 patients based on quantitative, operator-independent, lung CT densitometry

PURPOSE

To train and validate a predictive model of mortality for hospitalized COVID-19 patients based on lung densitometry.

METHODS

Two-hundred-fifty-one patients with respiratory symptoms underwent CT few days after hospitalization. "Aerated" (AV), "consolidated" (CV) and "intermediate" (IV) lung sub-volumes were quantified by an operator-independent method based on individual HU maximum gradient recognition. AV, CV, IV, CV/AV, IV/AV, and HU of the first peak position were extracted. Relevant clinical parameters were prospectively collected. The population was composed by training (n = 166) and validation (n = 85) consecutive cohorts, and backward multi-variate logistic regression was applied on the training group to build a CT_model. Similarly, models including only clinical parameters (CLIN_model) and both CT/clinical parameters (COMB_model) were developed. Model's performances were assessed by goodness-of-fit (H&L-test), calibration and discrimination. Model's performances were tested in the validation group.

RESULTS

Forty-three patients died (25/18 in training/validation). CT_model included AVmax (i.e. maximum AV between lungs), CV and CV/AE, while CLIN_model included random glycemia, C-reactive protein and biological drugs (protective). Goodness-of-fit and discrimination were similar (H&L:0.70 vs 0.80; AUC:0.80 vs 0.80). COMB_model including AVmax, CV, CV/AE, random glycemia, biological drugs and active cancer, outperformed both models (H&L:0.91; AUC:0.89, 95%CI:0.82-0.93). All models showed good calibration (R²:0.77-0.97). Despite several patient's characteristics were different between training and validation cohorts, performances in the validation cohort confirmed good calibration (R²:0-70-0.81) and discrimination for CT_model/COMB_model (AUC:0.72/0.76), while CLIN_model performed worse (AUC:0.64).

CONCLUSIONS

Few automatically extracted densitometry parameters with clear functional meaning predicted mortality of COVID-19 patients. Combined with clinical features, the resulting predictive model showed higher discrimination/calibration.

Automated CT Lung Density Analysis of Viral Pneumonia and Healthy Lungs Using Deep Learning-Based Segmentation, Histograms and HU Thresholds

CT patterns of viral pneumonia are usually only qualitatively described in radiology reports. Artificial intelligence enables automated and reliable segmentation of lungs with chest CT. Based on this, the purpose of this study was to derive meaningful imaging biomarkers reflecting CT patterns of viral pneumonia and assess their potential to discriminate between healthy lungs and lungs with viral pneumonia. This study used non-enhanced and CT pulmonary angiograms (CTPAs) of healthy lungs and viral pneumonia (SARS-CoV-2, influenza A/B) identified by radiology reports and RT-PCR results. After deep learning segmentation of the lungs, histogram-based and threshold-based analyses of lung attenuation were performed and compared. The derived imaging biomarkers were correlated with parameters of clinical and biochemical severity (modified WHO severity scale; c-reactive protein). For non-enhanced CTs (n = 526), all imaging biomarkers significantly differed between healthy lungs and lungs with viral pneumonia (all p < 0.001), a finding that was not reproduced for CTPAs (n = 504). Standard deviation (histogram-derived) and relative high attenuation area [600-0 HU] (HU-thresholding) differed most. The strongest correlation with disease severity was found for absolute high attenuation area [600-0 HU] (r = 0.56, 95% CI = 0.46-0.64). Deep-learning segmentation-based histogram and HU threshold analysis could be deployed in chest CT evaluation for the differentiating of healthy lungs from AP lungs.

Several specific high-resolution computed tomography patterns correlate with survival in patients with idiopathic pulmonary fibrosis

Background

Evidence of honeycombing in high-resolution computed tomography (HRCT) is a recognized risk factor for shortened survival in patients with idiopathic pulmonary fibrosis (IPF), but few studies have evaluated the feasibility of exploiting other specific patterns for predicting survival. The aim of this study was to examine the extent of specific HRCT patterns in IPF and determine whether they correlate with clinical features, pulmonary function tests (PFT), and survival.

Methods

Both the presence and extent of specific HRCT patterns, such as traction bronchiectasis, honeycombing, architectural distortion, reticulation, emphysema, and ground glass opacity, in 129 HRCT examinations were scored semi-quantitatively in three zones of each lung. HRCT examinations were also re-classified according to the 2011 and 2018 international statements. Correlations were calculated between the scores of specific HRCT patterns, clinical features, PFT, and patient survival.

Results

The extent of traction bronchiectasis was found to be an independent risk factor of shortened survival (HR 1.227, P=0.001). Patients with a possible usual interstitial pneumonia (UIP) pattern had a better median survival than the patients with a definite UIP pattern (61 vs. 37 months, P=0.026). The extents of traction bronchiectasis, honeycombing, and architectural distortion displayed an inverse correlation with all PFT values at the time of diagnosis. There were few differences between the radiological classifications of the 2011 and 2018 international statements.

Conclusions

We conclude that several specific HRCT patterns displayed a correlation with shortened survival in IPF; these may help in evaluating the risk of death in IPF patients.

[18F]FMISO PET/CT imaging of hypoxia as a non-invasive biomarker of disease progression and therapy efficacy in a preclinical model of pulmonary fibrosis: comparison with the [18F]FDG PET/CT approach

Purpose

Idiopathic pulmonary fibrosis (IPF) is a progressive disease with poor outcome and limited therapeutic options. Imaging of IPF is limited to high-resolution computed tomography (HRCT) which is often not sufficient for a definite diagnosis and has a limited impact on therapeutic decision and patient management. Hypoxia of the lung is a significant feature of IPF but its role on disease progression remains elusive. Thus, the aim of our study was to evaluate hypoxia imaging with [¹⁸F]FMISO as a predictive biomarker of disease progression and therapy efficacy in preclinical models of lung fibrosis in comparison with [¹⁸F]FDG.

Methods

Eight-week-old C57/BL6 mice received an intratracheal administration of bleomycin (BLM) at day (D) 0 to initiate lung fibrosis. Mice received pirfenidone (300 mg/kg) or nintedanib (60 mg/kg) by daily gavage from D9 to D23. Mice underwent successive PET/CT imaging at several stages of the disease (baseline, D8/D9, D15/D16, D22/D23) with [¹⁸F]FDG and [¹⁸F]FMISO. Histological determination of the lung expression of HIF-1α and GLUT-1 was performed at D23.

Results

We demonstrate that mean lung density on CT as well as [¹⁸F]FDG and [¹⁸F]FMISO uptakes are upregulated in established lung fibrosis (1.4-, 2.6- and 3.2-fold increase respectively). At early stages, lung areas with [¹⁸F]FMISO uptake are still appearing normal on CT scans and correspond to areas which will deteriorate towards fibrotic lesions at later timepoints. Nintedanib and pirfenidone dramatically and rapidly decreased mean lung density on CT as well as [¹⁸F]FDG and [¹⁸F]FMISO lung uptakes (pirfenidone: 1.2-, 2.9- and 2.6-fold decrease; nintedanib: 1.2-, 2.3- and 2.5-fold decrease respectively). Early [¹⁸F]FMISO lung uptake was correlated with aggressive disease progression and better nintedanib efficacy.

Conclusion

[¹⁸F]FMISO PET imaging is a promising tool to early detect and monitor lung fibrosis progression and therapy efficacy.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00259-021-05209-2.

Normal age-related quantitative CT values in the pediatric lung: from the first breath to adulthood

OBJECTIVE

To characterize the normal progression of quantitative CT parameters in normal children from birth to adulthood.

MATERIALS AND METHODS

Patients aged 0-18 years with non-contrast-enhanced chest CT and evidence of normal lung parenchyma were included. Patients with respiratory symptoms, incomplete anthropometric measurements, or sub-optimal imaging technique were excluded. Segmentation was performed using an open-source software with an automated threshold segmentation. The following parameters were obtained: mean lung density, kurtosis, skewness, lung volume, and mass. Linear and exponential regression models were calculated with age and height as independent variables. A p-value of <0.05 was considered significant.

RESULTS

220 patients (111 females, 109 males) were included. Mean age was 9.6 ± 5.9 years and mean height was 133.9 ± 35.1 cm. Simple linear regression showed a significant relationship between mean lung density with age (R 2 = 0.70) and height (R 2 = 0.73). Kurtosis displayed a significant exponential correlation with age (R 2 = 0.70) and height (R 2 = 0.71). Skewness showed a significant exponential correlation with age (R 2 = 0.71) and height (R 2 = 0.73). Lung mass showed a correlation with age (R 2 = 0.93) and height (R 2 = 0.92). Exponential regression showed a significant relationship between lung volume with age (R 2 = 0.88) and height (R 2 = 0.93).

CONCLUSION

Quantitative CT parameters of the lung parenchyma demonstrate changes from birth to adulthood. As children grow, the mean lung density decreases, and the lung parenchyma becomes more homogenous.

Comparative analysis of density histograms and visual scores in incremental and volumetric high-resolution computed tomography of the chest in idiopathic pulmonary fibrosis patients

Background

Volumetric high-resolution computed tomography (HRCT) of the chest has recently replaced incremental CT in the diagnostic workup of idiopathic pulmonary fibrosis (IPF). Concomitantly, visual and quantitative scores have been proposed for disease extent assessment to ameliorate disease management.

Purpose

To compare the performance of density histograms (mean lung attenuation, skewness, and kurtosis) and visual scores, along with lung function correlations, in IPF patients submitted to incremental or volumetric thorax HRCT.

Material and methods

Clinical data and CT scans of 89 newly diagnosed and therapy-naive IPF patients were retrospectively evaluated.

Results

Forty-six incremental and 43 volumetric CT scans were reviewed. No differences of density histograms and visual scores estimates were found by comparing two HRCT techniques, with an optimal inter-operator agreement (concordance correlation coefficient >0.90 in all instances). Single-breath diffusing lung capacity for carbon monoxide (DLCO_sb) was inversely related with the Best score (r = −00.416; p = 0.014), the Kazerooni fibrosis extent (r = −0.481; p = 0.004) and the mean lung attenuation (r = −0.382; p = 0.026), while a positive correlation was observed with skewness (r = 0.583; p = 0.001) and kurtosis (r = 0.543; p = 0.001) in the incremental HRCT sub-group. Similarly, in the volumetric CT sub-cohort, DLCO_sb was significantly associated with skewness (r = 0.581; p = 0.007) and kurtosis (r = 0.549; p = 0.018). Correlations with visual scores were not confirmed. Forced vital capacity significantly related to all density indices independently on HRCT technique.

Conclusions

Density histograms and visual scores similarly perform in incremental and volumetric HRCT. Density quantification displays an optimal reproducibility and proves to be superior to visual scoring as more strongly correlated with lung function.

Assessing the Effectiveness of Pirfenidone in Idiopathic Pulmonary Fibrosis: Long-Term, Real-World Data from European IPF Registry (eurIPFreg)

Background: Idiopathic pulmonary fibrosis (IPF) is a chronic progressive fibrotic pulmonary disease with rising incidence. In this study the effectiveness of pirfenidone, as measured by longitudinal change in individual slope of forced vital capacity (FVC) prior to and after initiating pirfenidone treatment, was evaluated in IPF patients recruited into the European registry for idiopathic pulmonary fibrosis (eurIPFreg). Secondary variables were the evaluation of the change in individual slope of diffusion capacity of the lungs for carbon monoxide (DLco), the Borg dyspnea scale, and six-minute walking distance (6MWD), as well as survival analyses. Results: Data of 122 eurIPFreg patients, who had at least two pulmonary function tests (PFTs) prior to or under treatment with pirfenidone, were analyzed by calculating slope-changes. The global analysis revealed an average slope change of +1.48 ± 0.28 (% per annum (p.a)) after start of treatment (p < 0.001), reflecting a reduction in annual FVC decline of approx. 50% under pirfenidone; it also showed a reduction in DLco, and increase in 6MWD (both p < 0.0001), as well as a flattening of the Borg dyspnea scale (p = 0.02). The median survival under treatment was 4.82 years. Patients with a more restrictive disease (FVC < 80% pred.), with a rapid progression (FVC decline >10% pred. p.a.), previous smokers and patients > 60 years of age seemed to profit more from pirfenidone treatment. Conclusions: We report the effectiveness of pirfenidone in a European “real world” IPF cohort with outcome data extending up to 9 years. Global analyses demonstrated a positive effect of pirfenidone on the decline of the lung function over time. Survival was dependent on Gender–Age–Physiology (GAP) score and age prior to therapy.

Objective quantitative multidetector computed tomography assessments in patients with combined pulmonary fibrosis with emphysema: Relationship with pulmonary function and clinical events

BACKGROUND

Combined pulmonary fibrosis with emphysema (CPFE) is a clinically meaningful syndrome characterized by coexisting upper-lobe emphysema and lower-lobe interstitial fibrosis. However, ambiguous diagnostic criteria and, particularly, the absence of objective methods to quantify emphysematous/fibrotic lesions in patients with CPFE confound the interpretation of the pathophysiology of this syndrome. We analyzed the relationship between objectively quantified computed tomography (CT) measurements and the results of pulmonary function testing (PFT) and clinical events in CPFE patients.

MATERIALS AND METHODS

We enrolled 46 CPFE patients who underwent CT and PFT. The extent of emphysematous lesions was obtained by calculating the percent of low attenuation area (%LAA). The extent of fibrotic lesions was calculated as the percent of high attenuation area (%HAA). %LAA and %HAA values were combined to yield the percent of abnormal area (%AA). We assessed the relationships between CT parameters and other clinical indices, including PFT results. Multivariate analysis was performed to examine the association between the CT parameters and clinical events.

RESULTS

A greater negative correlation with percent predicted diffusing capacity of the lung for carbon monoxide (DLCO %predicted) existed for %AA (r = -0.73, p < 0.001) than for %LAA or %HAA alone. The %HAA value was inversely correlated with percent predicted forced vital capacity (r = -0.48, p < 0.001), percent predicted total lung capacity (r = -0.48, p < 0.01), and DLCO %predicted (r = -0.47, p < 0.01). Multivariate logistic regression analysis found that %AA showed the strongest association with hospitalization events (odds ratio = 1.20, 95% confidence interval = 1.01-1.54, p = 0.029).

CONCLUSION

Quantitative CT measurements reflected deterioration in pulmonary function and were associated with hospitalization in patients with CPFE. This approach could serve as a useful method to determine the extent of lung morphology, pathophysiology, and the clinical course of patients with CPFE.

Clinical and Functional Characteristics of Patients with Unclassifiable Interstitial Lung Disease (uILD): Long-Term Follow-Up Data from European IPF Registry (eurIPFreg)

(1) Aim of the study: In spite of extensive research, up to 20% of interstitial lung diseases (ILD) patients cannot be safely classified. We analyzed clinical features, progression factors, and outcomes of unclassifiable ILD (uILD). (2) Methods: A total of 140 uILD subjects from the University of Giessen and Marburg Lung Center (UGMLC) were recruited between 11/2009 and 01/2019 into the European Registry for idiopathic pulmonary fibrosis (eurIPFreg) and followed until 01/2020. The diagnosis of uILD was applied only when a conclusive diagnosis could not be reached with certainty. (3) Results: In 46.4% of the patients, the uILD diagnosis was due to conflicting clinical, radiological, and pathological data. By applying the diagnostic criteria of usual interstitial pneumonia (UIP) based on computed tomography (CT), published by the Fleischner Society, 22.2% of the patients displayed a typical UIP pattern. We also showed that forced vital capacity (FVC) at baseline (p = 0.008), annual FVC decline ≥10% (p < 0.0001), smoking (p = 0.033), and a diffusing capacity of the lung for carbon monoxide (DLco) ≤55% of predicted value at baseline (p < 0.0001) were significantly associated with progressive disease. (4) Conclusions: The most important prognostic factors in uILD are baseline level and decline in lung function and smoking. The use of Fleischner diagnostic criteria allows further differentiation and accurate diagnosis.

Automatization and improvement of μCT analysis for murine lung disease models using a deep learning approach

Background

One of the main diagnostic tools for lung diseases in humans is computed tomography (CT). A miniaturized version, micro-CT (μCT) is utilized to examine small rodents including mice. However, fully automated threshold-based segmentation and subsequent quantification of severely damaged lungs requires visual inspection and manual correction.

Methods

Here we demonstrate the use of densitometry on regions of interest (ROI) in automatically detected portions of the lung, thus avoiding the need for lung segmentation. Utilizing deep learning approaches, the middle part of the lung is found in a μCT-stack and a ROI is placed in the left and the right lobe.

Results

The intensity values within the ROIs of the μCT images were collected and subsequently used for the calculation of different lung-related parameters, such as mean lung attenuation (MLA), mode, full width at half maximum (FWHM), and skewness. For validation, the densitometric approach was correlated with histological readouts (Ashcroft Score, Mean Linear Intercept).

Conclusion

We here show an automated tool that allows rapid and in-depth analysis of μCT scans of different murine models of lung disease.

Quantitative CT analysis for bronchiolitis obliterans in perinatally HIV-infected adolescents-comparison with controls and lung function data

OBJECTIVE

To compare quantitative chest CT parameters in perinatally HIV-infected adolescents with and without bronchiolitis obliterans compared with HIV-uninfected controls and their association with lung function measurements.

MATERIALS AND METHODS

Seventy-eight (41 girls) HIV-infected adolescents with a mean age of 13.8 ± 1.65 years and abnormal pulmonary function tests in the prospective Cape Town Adolescent Antiretroviral Cohort underwent contrast-enhanced chest CT on inspiration and expiration. Sixteen age-, sex-, and height-matched non-infected controls were identified retrospectively. Fifty-one HIV-infected adolescents (28 girls) displayed mosaic attenuation on expiration suggesting bronchiolitis obliterans. Pulmonary function tests were collected. The following parameters were obtained: low- and high-attenuation areas, mean lung density, kurtosis, skewness, ventilation heterogeneity, lung mass, and volume.

RESULTS

HIV-infected adolescents showed a significantly higher mean lung density, ventilation heterogeneity, mass, and high- and low-attenuation areas compared with non-infected individuals. Kurtosis and skewness were significantly lower as well. HIV-infected adolescents with bronchiolitis obliterans had a significantly lower kurtosis and skewness compared with those without bronchiolitis obliterans. Lung mass and volume showed the strongest correlations with forced expiratory volume in 1 s (FEV₁), forced vital capacity (FVC), and alveolar volume. Low-attenuation areas below - 950 HU and ventilation heterogeneity showed the strongest correlation with FEV₁/FVC (range, - 0.51 to - 0.34) and forced expiratory flow between 25 and 75% of FVC (range, - 0.50 to - 0.35).

CONCLUSION

Quantitative chest CT on inspiration is a feasible technique to differentiate perinatally HIV-infected adolescents with and without bronchiolitis obliterans. Quantitative CT parameters correlate with spirometric measurements of small-airway disease.

KEY POINTS

• Perinatally HIV-infected adolescents showed a more heterogeneous attenuation of the lung parenchyma with a higher percentage of low- and high-attenuation areas compared with non-infected patients. • Kurtosis and skewness are able to differentiate between HIV-infected adolescents with and without bronchiolitis obliterans using an inspiratory chest CT. • Quantitative CT parameters of the chest correlate significantly with pulmonary function test. Low-attenuation areas and ventilation heterogeneity are particularly associated with spirometric parameters related to airway obstruction.

Exploring the Ability of Electronic Nose Technology to Recognize Interstitial Lung Diseases (ILD) by Non-Invasive Breath Screening of Exhaled Volatile Compounds (VOC): A Pilot Study from the European IPF Registry (eurIPFreg) and Biobank

Background: There is an increasing interest in employing electronic nose technology in the diagnosis and monitoring of lung diseases. Interstitial lung diseases (ILD) are challenging in regard to setting an accurate diagnosis in a timely manner. Thus, there is a high unmet need in non-invasive diagnostic tests. This single-center explorative study aimed to evaluate the usefulness of electronic nose (Aeonose^®) in the diagnosis of ILDs. Methods: Exhaled volatile organic compound (VOC) signatures were obtained by Aeonose^® in 174 ILD patients, 23 patients with chronic obstructive pulmonary disease (COPD), and 33 healthy controls (HC). Results: By dichotomous comparison of VOC’s between ILD, COPD, and HC, a discriminating algorithm was established. In addition, direct analyses between the ILD subgroups, e.g., cryptogenic organizing pneumonia (COP, n = 28), idiopathic pulmonary fibrosis (IPF, n = 51), and connective tissue disease-associated ILD (CTD-ILD, n = 25) were performed. Area under the Curve (AUC) and Matthews’s correlation coefficient (MCC) were used to interpret the data. In direct comparison of the different ILD subgroups to HC, the algorithms developed on the basis of the Aeonose^® signatures allowed safe separation between IPF vs. HC (AUC of 0.95, MCC of 0.73), COP vs. HC (AUC 0.89, MCC 0.67), and CTD-ILD vs. HC (AUC 0.90, MCC 0.69). Additionally, to a case-control study design, the breath patterns of ILD subgroups were compared to each other. Following this approach, the sensitivity and specificity showed a relevant drop, which results in a poorer performance of the algorithm to separate the different ILD subgroups (IPF vs. COP with MCC 0.49, IPF vs. CTD-ILD with MCC 0.55, and COP vs. CT-ILD with MCC 0.40). Conclusions: The Aeonose^® showed some potential in separating ILD subgroups from HC. Unfortunately, when applying the algorithm to distinguish ILD subgroups from each other, the device showed low specificity. We suggest that artificial intelligence or principle compound analysis-based studies of a much broader data set of patients with ILDs may be much better suited to train these devices.

The therapy of idiopathic pulmonary fibrosis: what is next?

Idiopathic pulmonary fibrosis (IPF) is a chronic, progressive, fibrosing interstitial lung disease, characterised by progressive scarring of the lung and associated with a high burden of disease and early death. The pathophysiological understanding, clinical diagnostics and therapy of IPF have significantly evolved in recent years. While the recent introduction of the two antifibrotic drugs pirfenidone and nintedanib led to a significant reduction in lung function decline, there is still no cure for IPF; thus, new therapeutic approaches are needed. Currently, several clinical phase I-III trials are focusing on novel therapeutic targets. Furthermore, new approaches in nonpharmacological treatments in palliative care, pulmonary rehabilitation, lung transplantation, management of comorbidities and acute exacerbations aim to improve symptom control and quality of life. Here we summarise new therapeutic attempts and potential future approaches to treat this devastating disease.

Clinical characteristics of patients with familial idiopathic pulmonary fibrosis (f-IPF)

Background

The aim of this study was to analyze the relative frequency, clinical characteristics, disease onset and progression in f-IPF vs. sporadic IPF (s-IPF).

Methods

Familial IPF index patients and their family members were recruited into the European IPF registry/biobank (eurIPFreg) at the Universities of Giessen and Marburg (UGMLC). Initially, we employed wide range criteria of f-IPF (e.g. relatives who presumably died of some kind of parenchymal lung disease). After narrowing down the search to occurrence of idiopathic interstitial pneumonia (IIP) in at least one first grade relative, 28 index patients were finally identified, prospectively interviewed and examined. Their family members were phenotyped with establishment of pedigree charts.

Results

Within the 28 IPF families, overall 79 patients with f-IPF were identified. In the same observation period, 286 f-IIP and s-IIP patients were recruited into the eurIPFreg at our UGMLC sites, corresponding to a familial versus s-IPF of 9.8%. The both groups showed no difference in demographics (61 vs. 79% males), smoking history, and exposure to any environmental triggers known to cause lung fibrosis. The f-IPF group differed by an earlier age at the onset of the disease (55.4 vs. 63.2 years; p < 0.001). On average, the f-IPF patients presented a significantly milder extent of functional impairment at the time point of inclusion vs. the s-IPF group (FVC 75% pred. vs. FVC 62% pred., p = 0.011). In contrast, the decline in FVC was found to be faster in the f-IPF vs. the s-IPF group (4.94% decline in 6 months in f-IPF vs. 2.48% in s-IPF, p = 0.12). The average age of death in f-IPF group was 67 years vs. 71.8 years in s-IPF group (p = 0.059). The f-IIP group displayed diverse inheritance patterns, mostly autosomal-dominant with variable penetrance. In the f-IPF, the younger generations showed a tendency for earlier manifestation of IPF vs. the older generation (58 vs. 66 years, p = 0.013).

Conclusions

The 28 f-IPF index patients presented an earlier onset and more aggressive natural course of the disease. The disease seems to affect consecutive generations at a younger age.

Trial registration

Nr. NCT02951416http://www.www.clinicaltrials.gov

Exhalative Breath Markers Do Not Offer for Diagnosis of Interstitial Lung Diseases: Data from the European IPF Registry (eurIPFreg) and Biobank

Background: New biomarkers are urgently needed to facilitate diagnosis in Interstitial Lung Diseases (ILD), thus reducing the need for invasive procedures, and to enable tailoring and monitoring of medical treatment. Methods: In this study we investigated if patients with idiopathic pulmonary fibrosis (IPF; n = 21), non-IPF ILDs (n = 57) and other lung diseases (chronic obstructive pulmonary disease (COPD) n = 24, lung cancer (LC) n = 16) as well as healthy subjects (n = 20) show relevant differences in exhaled NO (FeNO; Niox MINO), or in eicosanoid (PGE2, 8-isoprostane; enzyme-linked immunosorbent assay (ELISA)) levels as measured in exhaled breath condensates (EBC) and bronchoalveolar lavage fluids (BALF). Results: There was no significant difference in FeNO values between IPF, non-IPF ILDs and healthy subjects, although some individual patients showed highly elevated FeNO. On the basis of the FeNO signal, it was neither possible to differentiate between the kind of disease nor to detect exacerbations. In addition, there was no correlation between FeNO values and lung function. The investigation of the eicosanoids in EBCs was challenging (PGE2) or unreliable (8-isoprostane), but worked out well in BALF. A significant increase of free 8-isoprostane was observed in BALF, but not in EBCs, of patients with IPF, hypersensitivity pneumonitis (HP) and sarcoidosis, possibly indicating severity of oxidative stress. Conclusions: FeNO-measurements are not of diagnostic benefit in different ILDs including IPF. The same holds true for PGE2 and 8-isoprostane in EBC by ELISA.